* [Drbd-dev] cluster initialization with drbdmanage
@ 2015-05-20 9:26 Dietmar Maurer
2015-05-20 11:55 ` Roland Kammerer
0 siblings, 1 reply; 7+ messages in thread
From: Dietmar Maurer @ 2015-05-20 9:26 UTC (permalink / raw)
To: drbd-dev
Hi all,
The drbdmanage installation HowTo (README) contains the following instaructions:
>Cluster initialization (first drbdmanage cluster node only):
> * On the first node, issue the following command to initialize a new
> drbdmanage cluster:
> drbdmanage init <ipaddress>
This works. Now I want to add the second node:
>Adding nodes to a drbdmanage cluster:
> * On a node of the existing drbdmanage cluster, issue the following
> command to prepare the cluster for integrating a new node:
> drbdmanage new-node <nodename> <ipaddress>
> ... where <nodename> is the node name of the node, and <ipaddress>
> is the IP address of the node that will join the cluster
> This command will also generate the command line that needs to be
> entered on the new node to make it join the cluster
Now I run:
# drbdmanage new-node hatest2 192.168.3.202
Operation completed successfully
Operation completed successfully
Executing join command on hatest2 using ssh.
You are going to join an existing drbdmanage cluster.
CAUTION! Note that:
* Any previous drbdmanage cluster information may be removed
* Any remaining resources managed by a previous drbdmanage installation
that still exist on this system will no longer be managed by drbdmanage
Confirm:
yes/no: yes
Logical volume ".drbdctrl" successfully removed
Logical volume ".drbdctrl" created.
You want me to create a v09 style flexible-size internal meta data block.
There appears to be a v09 flexible-size internal meta data block
already in place on /dev/drbdpool/.drbdctrl at byte offset 4190208
Do you really want to overwrite the existing meta-data?
*** confirmation forced via --force option ***
Do you want to proceed?
*** confirmation forced via --force option ***
NOT initializing bitmap
md_offset 4190208
al_offset 4157440
bm_offset 4153344
Found some data
==> This might destroy existing data! <==
initializing activity log
Writing meta data...
New drbd meta data block successfully created.
Operation completed successfully
--------------------
I have two questions now:
1.) Where is the command line to join the node (mentioned in the README above)?
2.) Does this really destroy all existing resources? If so, how can I add
nodes to a cluster which already have existing resources?
Or is there another way to initialize the cluster?
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Drbd-dev] cluster initialization with drbdmanage
2015-05-20 9:26 [Drbd-dev] cluster initialization with drbdmanage Dietmar Maurer
@ 2015-05-20 11:55 ` Roland Kammerer
2015-05-20 12:09 ` Dietmar Maurer
0 siblings, 1 reply; 7+ messages in thread
From: Roland Kammerer @ 2015-05-20 11:55 UTC (permalink / raw)
To: drbd-dev
On Wed, May 20, 2015 at 11:26:27AM +0200, Dietmar Maurer wrote:
> I have two questions now:
>
> 1.) Where is the command line to join the node (mentioned in the README above)?
If the node where you execute "add-node" is able to ssh to the new node,
everything is fine and the join is executed for you. If not, the join
command gets printed and you have to manually execute the command on the
new node.
If you want to know which command that would be, you can use:
drbdmanage howto-join <nodename>
> 2.) Does this really destroy all existing resources? If so, how can I add
> nodes to a cluster which already have existing resources?
Resouces managed by drbdmanage or other drbd (non drbdmanage manged)
resources?
Regards, rck
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Drbd-dev] cluster initialization with drbdmanage
2015-05-20 11:55 ` Roland Kammerer
@ 2015-05-20 12:09 ` Dietmar Maurer
2015-05-20 12:40 ` Roland Kammerer
0 siblings, 1 reply; 7+ messages in thread
From: Dietmar Maurer @ 2015-05-20 12:09 UTC (permalink / raw)
To: Roland Kammerer, drbd-dev
> > 1.) Where is the command line to join the node (mentioned in the README
> > above)?
>
> If the node where you execute "add-node" is able to ssh to the new node,
> everything is fine and the join is executed for you.
The README tell to execute new-node always on
"a node of the existing drbdmanage cluster"
So for example, I should run 'add-node' always on the first node?
Is that correct?
> If not, the join
> command gets printed and you have to manually execute the command on the
> new node.
>
> If you want to know which command that would be, you can use:
> drbdmanage howto-join <nodename>
>
> > 2.) Does this really destroy all existing resources? If so, how can I add
> > nodes to a cluster which already have existing resources?
>
> Resouces managed by drbdmanage
yes
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Drbd-dev] cluster initialization with drbdmanage
2015-05-20 12:09 ` Dietmar Maurer
@ 2015-05-20 12:40 ` Roland Kammerer
2015-05-20 12:54 ` Dietmar Maurer
0 siblings, 1 reply; 7+ messages in thread
From: Roland Kammerer @ 2015-05-20 12:40 UTC (permalink / raw)
To: drbd-dev
On Wed, May 20, 2015 at 02:09:03PM +0200, Dietmar Maurer wrote:
> > > 1.) Where is the command line to join the node (mentioned in the README
> > > above)?
> >
> > If the node where you execute "add-node" is able to ssh to the new node,
> > everything is fine and the join is executed for you.
>
> The README tell to execute new-node always on
> "a node of the existing drbdmanage cluster"
> So for example, I should run 'add-node' always on the first node?
> Is that correct?
On any node that is already in the cluster... I always use the first
node.
>
> > If not, the join
> > command gets printed and you have to manually execute the command on the
> > new node.
> >
> > If you want to know which command that would be, you can use:
> > drbdmanage howto-join <nodename>
> >
> > > 2.) Does this really destroy all existing resources? If so, how can I add
> > > nodes to a cluster which already have existing resources?
> >
> > Resouces managed by drbdmanage
>
> yes
>
In that case it is exactly as drbdmanage told you:
CAUTION! Note that:
* Any previous drbdmanage cluster information may be removed
* Any remaining resources managed by a previous drbdmanage installation
that still exist on this system will no longer be managed by drbdmanage
The control volume gets overwritten, the old information is lost.
Regards, rck
P.S.:
1) Don't TO/CC me, I'm on the list.
2) Use the drbd-user list for that kind of questions.
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Drbd-dev] cluster initialization with drbdmanage
2015-05-20 12:40 ` Roland Kammerer
@ 2015-05-20 12:54 ` Dietmar Maurer
2015-05-20 13:23 ` Roland Kammerer
0 siblings, 1 reply; 7+ messages in thread
From: Dietmar Maurer @ 2015-05-20 12:54 UTC (permalink / raw)
To: Roland Kammerer, drbd-dev
> In that case it is exactly as drbdmanage told you:
>
> CAUTION! Note that:
> * Any previous drbdmanage cluster information may be removed
> * Any remaining resources managed by a previous drbdmanage installation
> that still exist on this system will no longer be managed by drbdmanage
>
> The control volume gets overwritten, the old information is lost.
But what if I want to add additional nodes to a cluster later?
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Drbd-dev] cluster initialization with drbdmanage
2015-05-20 12:54 ` Dietmar Maurer
@ 2015-05-20 13:23 ` Roland Kammerer
2015-05-20 15:02 ` Dietmar Maurer
0 siblings, 1 reply; 7+ messages in thread
From: Roland Kammerer @ 2015-05-20 13:23 UTC (permalink / raw)
To: drbd-dev
On Wed, May 20, 2015 at 02:54:54PM +0200, Dietmar Maurer wrote:
> > In that case it is exactly as drbdmanage told you:
> >
> > CAUTION! Note that:
> > * Any previous drbdmanage cluster information may be removed
> > * Any remaining resources managed by a previous drbdmanage installation
> > that still exist on this system will no longer be managed by drbdmanage
> >
> > The control volume gets overwritten, the old information is lost.
>
> But what if I want to add additional nodes to a cluster later?
It is fine to add additional nodes later as long as they are really new
and do not contain a control volume or you are willing to overwrite the
existing control volume of that node. The state of your cluster is not
affected by that, only the new node.
I guess, what you mean by "add additional nodes" is merging two
clusters, right? So adding a node node X to cluster A, that was
previously in cluster B and contains resources from the old cluster B.
Right? If it would not have been part of some kind of cluster B, it
would not have a control volume. "add-node" is synonymous for
"new-node". And "new-node" means that this node becomes part of the new
cluster and its old information is not required anymore.
It is the third time I ask you to switch to drbd-user and to not TO/CC
me. Is it really that difficult?
Regards, rck
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [Drbd-dev] cluster initialization with drbdmanage
2015-05-20 13:23 ` Roland Kammerer
@ 2015-05-20 15:02 ` Dietmar Maurer
0 siblings, 0 replies; 7+ messages in thread
From: Dietmar Maurer @ 2015-05-20 15:02 UTC (permalink / raw)
To: drbd-dev
> It is the third time I ask you to switch to drbd-user and to not TO/CC
> me. Is it really that difficult?
Yes, it is difficult because:
1.) I consider it bad behavior to switch an existing thread
from one list to another. The archives gets inconsistent that way.
2.) It is unclear what you mean by 'such questions'. For me all this
are development related issues (bugs, missing features, ...).
3.) My environment adds your email address automatically if I press reply.
As requested, I will post further replies to drbd-user.
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2015-05-20 15:02 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-05-20 9:26 [Drbd-dev] cluster initialization with drbdmanage Dietmar Maurer
2015-05-20 11:55 ` Roland Kammerer
2015-05-20 12:09 ` Dietmar Maurer
2015-05-20 12:40 ` Roland Kammerer
2015-05-20 12:54 ` Dietmar Maurer
2015-05-20 13:23 ` Roland Kammerer
2015-05-20 15:02 ` Dietmar Maurer
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.